Error: object 'var' not found when running analysis inside a function

[alejandroglez]

Hi there,
I have a data frame in a long format, containing multiple records by individual, corresponding to multiple eating occasions (e.g breakfast, snack, etc.) along with sodium intake in each occasion (see data set below). Then I convert it to a long format, so that each eating occasion is a different variable. Finally, I compute the mean of sodium by breakfast. See code below:
library(tidyr)
mydat = read.csv(file="…MyFile.csv")
#Transposing form long to wide format
mydat.SOD = as.data.frame(pivot_wider(mydat,
names_from = food.occasion,
values_from = SOD))
mydat.SOD[is.na(mydat.SOD)] = 0
#Renaming
names(mydat.SOD)[names(mydat.SOD) == "Breakfast"] = "SOD.brf"
names(mydat.SOD)[names(mydat.SOD) == "Lunch/Brunch"] = "SOD.lun"
names(mydat.SOD)[names(mydat.SOD) == "Dinner"] = "SOD.din"
names(mydat.SOD)[names(mydat.SOD) == "Snack"] = "SOD.snk"
names(mydat.SOD)[names(mydat.SOD) == "Drink"] = "SOD.drk"
mean(mydat.SOD$SOD.brf)
This code works perfectly. However, when I create a function to run this code multiple times (i.e. other nutrients), the code give me an error. Please, see the code below.
myF = function(nutrient){
#Transposing form long to wide format
mydat.nutrient = as.data.frame(pivot_wider(mydat,
names_from = food.occasion,
values_from = nutrient))
mydat.nutrient[is.na(mydat.nutrient)] = 0

#Renaming
names(mydat.nutrient)[names(mydat.nutrient) == "Breakfast"] = "nutrient.brf"
names(mydat.nutrient)[names(mydat.nutrient) == "Lunch/Brunch"] = "nutrient.lun"
names(mydat.nutrient)[names(mydat.nutrient) == "Dinner"] = "nutrient.din"
names(mydat.nutrient)[names(mydat.nutrient) == "Snack"] = "nutrient.snk"
names(mydat.nutrient)[names(mydat.nutrient) == "Drink"] = "nutrient.drk"

mean(mydat.nutrient$nutrient.brf)
}

myF(SOD)
Error: object 'SOD' not found
Run rlang::last_error() to see where the error occurred.

Can you please advise?
Thank you.

Please, see the data set I'm using below.

ID	food.occasion	SOD
1	Breakfast	248.8
1	Lunch/Brunch	707.6
1	Dinner	424.8
1	Snack	2.3
1	Drink	11.7
2	Breakfast	824.7
2	Lunch/Brunch	511.8
2	Dinner	1813.3
2	Snack	166.8
2	Drink	102.2
3	Lunch/Brunch	1108.1
3	Dinner	179.6
3	Drink	19.6
4	Breakfast	332.8
4	Lunch/Brunch	1030.9
4	Dinner	109
5	Breakfast	300.1
5	Lunch/Brunch	150.6
5	Dinner	1896.9
5	Drink	23.4

Additional relevant information:
OS: Windows 10 (64-bit)
R version: 3.6.2
R studio version: 3.5

[nirgrahamuk]

myF = function(nutrient){
  nutrient <- rlang::ensym(nutrient)
  #Transposing form long to wide format
  mydat.nutrient = as.data.frame(pivot_wide...

[alejandroglez]

Hi nirgrahamuk,
Thanks a lot for your quick response. I worked with no error. However, when the input data frame contains more than one nutrient (which is the case in my actual data), the code does not produce the expected results (see new data with an additional nutrient. Instead of getting one observation per individual, the output file retains all the original records and the transposed values appear in some kind of diagonal pattern. I was trying to subset the input data instead within the “pivot_wider” function, but I get errors.

subset(mydat,select=c(SAMPLEID,nutrient,food.occasion),

Additionally, the output data frame contains in the name of the variables the word “nutrient” (e.g. nutrient.brf, nutrient.lun, etc.) instead of SOD.bfr,…This is not convenient to me, since I'll run that function for many nutrients, and them combine all the resulting data frames. Therefore, I don't want having variables with the same name.
Thanks again,
A.G.

ID	food.occasion	SOD	SUG
1	Breakfast	248.8	0.9
1	Lunch/Brunch	707.6	18.5
1	Dinner	424.8	27.2
1	Snack	2.3	3.8
1	Drink	11.7	0
2	Breakfast	824.7	11.6
2	Lunch/Brunch	511.8	14.5
2	Dinner	1813.3	22.6
2	Snack	166.8	0
2	Drink	102.2	16.9
3	Lunch/Brunch	1108.1	0
3	Dinner	179.6	0.3
3	Drink	19.6	12.5
4	Breakfast	332.8	11.8
4	Lunch/Brunch	1030.9	16.8
4	Dinner	109	33.5
5	Breakfast	300.1	0
5	Lunch/Brunch	150.6	9.6
5	Dinner	1896.9	3
5	Drink	23.4	0